2.7.5

• Overlay of detected points in original image
• Caerus import fix
• added alpha for plotting detected points in image
• make sure the order is correct for plotting
• add try catch for div by 0
• read image as unchanged and thus do not force into RGB color image
• chain fix for pandas
• in case image is already gray scale, do nothing
• some code refactoring and cleaning

2.7.4

• Fixes to let it work for colab notebook

2.7.3

It is surprising where bugs can hide. A few were hidden due to the numpy upgrade and new parameters.

• bugfix for make unique when 1D vector or 2D array.
• Bug fixes due to new numpy with floats and uint8
• bugfix to shown text labels in 1D and 2D images

2.7.2

• Fix for showing text in the plot_persistence
• Added new parameters for plot_persistence: s=20, marker='x', color='#FF0000'
• Reset index for persistence in results: results['persistence'].reset_index(inplace=True, drop=True)

2.7.1

• Optimized functionality (make_unique) because it was slow.

2.7.0

A large update!

• Restriction removed for Numpy (version 2.x) can now be installed #38
• Fix for peak detection with topological method of equal height #39
• Fix in case fastnl or bilateral is used. Then, scaling must be done too.
• Contribution of @martinuray is included for neighborhood extraction [1]
• New functionality for plotting: params_line, params_peak_marker and params_valley_marker is added as parameter which allows you to customize your line and detection markers
• Docstrings updated
• Examples updated
• Some cleaning and code refactoring

[1] J. Ferner, S. Huber, S. Messineo, A. Pop, and M. Uray, “Persistence-based Hough Transform for Line Detection,” May 22, 2025, arXiv: arXiv:2504.16114. doi: 10.48550/arXiv.2504.16114

2.6.6

• Updates to support pandas 3.0
• Updates in requirements to prevent issues
• update in init file
• Use incremental adjustments instead of random noise to avoid duplicated data

Also thanks to @dstrain115 @kohlert

v2.6.5

• Improved implementation finding peaks in cases with small absolute variance #26 #35 #36
  Thank you @kohlert

2.6.4

• Bug is fixed in peakdetect
• Peakdetect 1.2 is used
• Newest scipy is used again
• Fix for bug in caerus

v2.6.3

• code refactoring and cleaning by @arvinnick
  Thank you for the contribution!